An Essay Replete With Titillating Narrative Carefully Selected to Illustrate the Efficacy, Practice and Application, as well as the Pitfalls, of Modern Radiocarbon Dating Techniques, Accompanied by Numerous Illustrations, Important Cautionary Tales for the Archaeologist and Several Mercifully Brief Anecdotes from the Monterey Bay Area
|Paper presented at the Annual Meeting of the Society for California Archaeology, San Diego, California, April 6, 2002.|
Click here to download a pdf copy of this report (144 k).
Abstract: Although radiocarbon dating has been practiced in California for 50 years, many archaeologists are still not realizing the full potential of the technique. Some of the problems we have identified in recent projects are: selecting samples which fail to adequately characterize the full temporal range of a site; employing too few samples to accurately date a site; and failure to report one or more of the pieces of information necessary for others to understand the results of the radiocarbon dating.
This is a story about dating. Unfortunately, much of this story is about dating gone wrong. The rest is how dating techniques can be improved.
Radiocarbon dating has been practiced in California for 50 years, but many archaeologists still do not realize the full potential of the technique.
This paper addresses the coast of Central California, where virtually all sites have adequate amounts of shell or other materials for dating. On the Monterey Peninsula, virtually all sites contain quantities of abalone and mussel shell, so these are emphasized in the following paper.
In some parts of California, radiocarbon dating is more difficult because of the lack of adequate samples, and some of the following comments will apply to a lesser degree to those areas.
We have identified several problems which are rampant in central California archaeology. These include:
Two of these three problems are illustrated in the following "Cautionary Tale."
- 1. Selecting samples which fail to characterize the full temporal range of a site;
- 2. Employing too few samples to accurately date a site and its various components; and
- 3. Failure to report one or more of the pieces of information necessary for others to understand the results of a dating project. This is particularly critical information for future researchers to be able to reinterpret our data based on new information.
Cautionary Tale 1 -- Dating 103
CA-MNT-103 that is, a site located on the eastern end of Cannery Row in Monterey. This site, along with probably all the other sites on Cannery Row is a portion of a single large occupation area used serially or intermittently for somewhere over 5,000 years.
In order to compare the common practice of dating mixed shell samples against dating single-shell samples  we conducted an experiment. Although we did not date the site in this way, our Cautionary Tale serves to illustrate two of the problems noted previously.
One of the most common methods of dating is to obtain a single sample from a deposit. Commonly, this is a mixed-material sample. At CA-MNT-103, we obtained a sample from the exact middle of the deposit using a handful of small mussel shell fragments (324 to be exact; see Figure 1).
Figure 1. Shell Sample Used for Radiocarbon Dating at CA-MNT-103.
We dated this sample, and got the result shown in Figure 2. The calibrated date came back at 1090 B.C., at the end of the Middle Period as recognized on the Monterey Peninsula.
Knowing that a single date is not enough, we dated a couple of large shell fragments from an abalone layer between 20 and 30 cm. This returned dates of A.D. 1240 and 1260, evidence of a Late Period component. Great! We now have two components! (Figure 3).
Next we submitted six single-piece mussel shell samples weighing between 0.4 and 1.5 g for AMS dating. This confirmed the Late Period component! Look at the tight range for all of these samples (Figure 4).
Finally, we submitted nine single-piece mussel shell samples for AMS dating, bringing the total number of samples to 18. Look! Another Early Period component! (Figure 5). But this one is earlier than the first! What's going on?
Actually, as you have probably guessed, we didn't submit these samples in the order described above. The 17 single-shell samples were part of a tightly controlled experiment and were all submitted at the same time. They showed a very narrow component about A.D. 1175 and a second component about 2140 B.C.
Subsequently we submitted the sample of 324 pieces of mussel shell from the 30-40 cm level to explore the difference between single-shell and multiple-shell dates. And as has been shown, that sample has no basis in fact (Figure 6).
Even with 17 single-shell radiocarbon dates we can't guarantee to have established the full temporal range of even this portion of the deposit. We have only two samples of abalone shell, and these are the two youngest samples. Also, there may be abalone shells associated with the Early Period deposit which have not been dated.
Figure 6. The Actual Dates from CA-MNT-103.
However, we have been able to establish several important points concerning CA-MNT-103 and archaeological radiocarbon dating on the Monterey Peninsula:
- These rigorous dating techniques showed that there were two intervals during which this portion of the site was occupied, one during the Early Period and one during the Late Period. Each lasted for a relatively short time.
- Bioturbation has resulted in the thorough intermixing of mussel shell fragments from these two occupation periods. In this area of the site there is no statistical correlation between the age of a mussel shell fragment and its depth.
- The abalone layer between 20 and 30 cm appears undisturbed, and associates closely with, but slightly younger than, the Late Period mussel shell dates. Dating abalone shells from this layer will not characterize the full range of occupation of either the Late Period component or this area of the site.
- Mussel shell fragments may accurately characterize the two different occupation intervals at this site if a sufficiently large number of samples is employed. We suspect that the 15 mussel shell samples we obtained is close to sufficient, but only a much larger sample can determine this conclusively. One problem we see is that the two abalone shell samples are both younger than the six Late Period mussel shell samples.
- There have been at least three tests on the Monterey Peninsula which compared abalone and mussel shells supposedly of the same age. In each case, abalone shells dated approximately 100-110 years younger than the corresponding mussel shells. It is possible that there are interspecies or cultural differences affecting these samples, but in any case this phenomenon needs to be explored further.
- If you date a series of features, you date specific cultural activities, but those features do not necessarily date the full range of the site. The abalone layer between 20-30 cm at CA-MNT-103 does not accurately date the Late Period component, and fails miserably to date the Early Period component of the site.
You can see from this Cautionary Tale two problems are that are common in the literature. First, poor sample selection (especially mixed shell samples) can give inaccurate results. Second, dating too few samples can give incomplete results.
On the Monterey Peninsula shell is an ideal material for radiocarbon dating, but in addition to the normal problems inherent in its use, we have identified several additional problems:
- Multiple-shell and bulk soil samples can give inaccurate results, and should be used with caution or discontinued entirely.
- The good news: large pieces of shell, such as abalone, are less likely to be shifted about within a deposit by bioturbation than are smaller items, such as mussel shell fragments. Large pieces of shell are also cheaper to date.
- The bad news: abalone shells are more likely to cluster into discrete layers representing limited temporal periods and activities. Dating only the abalone shells may accurately date a Late Period abalone processing site, but will probably not date the full range of occupation of a multicomponent site.
- Mussel shell fragments are found in all temporal periods in the Monterey Bay, although they may not be present in some specialized sites. Owing to the ease with which they can be broken and moved, mussel shells are less likely to retain their original provenience; a date on a fragment of mussel shell from the top of a site may actually date a lower component.
- Red abalone shells from Late Period sites average about 200 years younger than mussel shells. 
- Red abalone shells from Middle Period sites average about 300 years younger than mussel shells. 
HOW MANY SAMPLES DO YOU NEED?
Even if you select your samples carefully, how many samples is enough? With many archaeological sites, the only way you can tell if you have obtained enough samples is by obtaining too many samples. Not good for your budget, but good for science!
- To accurately date a site of the kind we find in the Monterey Bay area, it is necessary to use single-shell samples and to date both the abalone layer (if present) and a representative sample of mussel shells. It is equally important to date a sufficiently large number of samples to characterize the full range of the different components which may be present. There will frequently be gaps between components, so the number of samples needs to be high enough to characterize these as well. In looking at the literature, most complex sites which have been sampled probably contain occupational and abandonment episodes which the dating strategies failed to identify.
- The number of dates required to characterize a complex site can't be determined in advance; it can only be determined by obtaining a large enough suite of dates so that you know you have too many.
- The common practice of submitting a few samples to the lab and writing up the results as soon as the numbers come back may not be adequate. You may need to submit more samples to figure out what's really going on. And you may need to do this three or four times to get it right!
- Not all parts of a site will date to the same time period or periods. Intrasite variation is one of the least studied and least understood problems still facing radiocarbon dating in Central California.
REPORTING THE RESULTS
There are many examples of poorly reported archaeological dates in the literature. We have seen examples where radiocarbon dates did not include one or more of the following: the laboratory number, the measured age, the species of shell used, the number of pieces of shell used, whether or not a C13/C12 ratio was obtained, and other critical data needed to understand the date.
When you cite a date, what are you citing? Is the date measured age, conventional age, or calibrated age? If calibrated age, what Delta-R did you use?
Finally, what do you as the author think about the accuracy of the date, and what do you think the date means in terms of the people who occupied the site?
For each date it is critical for the reader of a report to be able to ascertain the following:
- The precise nature of your sample, including the material you used, the number of separate pieces, the weight, and the exact provenience.
- The method of analysis, and
- The results of the analysis, including all the data sheets from the radiocarbon dating laboratory.
Cautionary Tale 2 -- CA-MNT-234, Moss Landing
This large site, with a total of 66 radiocarbon dates is one of the best-dated sites in California, but only 29 dates are on single-shell samples. Three additional dates are on single pieces of bone.
Figure 8 (lower half) shows all 66 radiocarbon dates. They make a nice smooth pattern, showing a generally continuous occupation of nearly 8,000 years (measured age).
But if you leave out the bulk soil, multiple-shell, and miscellaneous charred material (34 dates) you get an entirely different picture (Figure 8, upper). Notice the 2,320 year gap during which the site was possibly abandoned. There are three other gaps whose meanings are less clear. They could be periods of abandonment, but at some point increasingly smaller gaps are more likely to be an artifact of the radiocarbon dating strategy.
Figure 8. Radiocarbon Dates from CA-MNT-234.
Lower: all 66 dates. Upper: Single shell and bone dates.
This is an example where additional single-shell dates would clarify the occupational periods and the abandonments at this site. The three single-piece bone dates were on marine mammals which when calibrated agreed closely with the shell dates.
Multiple-shell and bulk soil samples mix a lot of individual things together. Dating "many old things" is no more accurate than dating "any old thing." This intermixing obscures significant data
Cautionary Tale 3 -- Monterey County
When you examine the nearly 600 radiocarbon dates for Monterey County, you get entirely different views of the area's prehistory when you consider single-shell dates (Figure 9, upper) vs. the full database (Figure 9, lower): there are gaps between the Early, Middle, and Late periods which do not appear in the full database (these are shown in more detail in Figure 10).
Figure 9. Radiocarbon Dates from Monterey County. Lower: nearly 600 dates.
Upper: Single shell dates. The area highlighted in yellow appears in Figure 10.
Figure 10. Single-shell Radiocarbon Dates from Monterey County, 1150-3450 B.P.
Note the gaps between the Early, Middle, and Late Periods.
Something is going on with these gaps--this is real information--but it only shows up in the single-shell database.
- At what has been called the Middle/Late Transition there are two gaps of 60 and 130 years, or a span of 190 years containing only one date. 
- At what has been called the Early/Middle Transition there are two gaps of 265 and 245 years, or a span of 510 years containing only one date. If this is expanded, there is a gap of 800 years containing only three dates. 
The single-shell database shows significant gaps not shown by the full database. Some of the gaps appear to be culturally significant.
Cautionary Tale 4 -- CA-MNT-437, Carmel Highlands
This is another interesting site where the age depends on which shellfish species you date. If you date mixed shell (the two black bars on the left) you don't have a clue. The mussel shell dates (blue bars) give a pretty good idea of the overall age of the deposit, but the abalone shell dates (red bars) miss the Middle Period component.
Figure 11. Radiocarbon Dates from CA-MNT-437.
Abalone shells date only the upper component. Mussel shells date both components. Would more samples identify a third component? Were enough samples obtained?
Cautionary Tale 5 -- CA-MNT-820, Carmel Highlands
This is another example where the first two dates on mixed shell (the two black bars on the left) provided false information. Additional dates using single pieces of abalone shell provided better information. We can only guess what the results might have been if mussel shells were dated as well.
Figure 12. Radiocarbon Dates from CA-MNT-820.
Abalone shells identify both a Late and an Early component. Would mussel shells have provided additional information?
To summarize, in coastal Central California, and possibly other areas of California, we are not yet obtaining as much information from radiocarbon dating as we should. We have a powerful dating tool, if we use it correctly and adequately.
Overall, we are:
- Processing too few samples;
- Selecting samples which provide only limited data;
- Failing to accurately report the results of the samples we do obtain; and
- Failing to be analytically critical of our database.
Fifty years after the introduction of radiocarbon dating it is time to start doing a better job.
 Even Santa Barbara County, one of the best dated counties in California, had over a third of its sites dated with a single date in 1996 (Breschini et al. 1996). Return
 Measured age 784 vs. 984 years B.P.; 132 abalone shell samples and 33 mussel shell samples. Return
 Measured age 1958 vs. 2253 years B.P.; 14 abalone shell samples and 7 mussel shell samples. The sample from the Early Period is currently too limited to reach any conclusions. Return
 The Late Period has an average separation between dates of 6.1 years. The largest gap within this period is 40 years. Return
 During the Middle Period the separation between dates averages 41.3 years. Return
|Entire contents Copyright 2001-2011 by Coyote Press. All rights reserved. CALIFORNIAPREHISTORY.COM, CALIFORNIAPREHISTORY.NET, CALIFORNIAPREHISTORY.ORG, and the CALIFORNIAPREHISTORY MASTHEAD are all trademarks of Coyote Press, Gary S. Breschini, and Trudy Haversat.||Back to the Top|